Welcome back to The Battery Buildout.
Below is the frame for our “Battery Bigafactory Building Practice” series:
Part 1 | Part 2 | Part 3 | Part 4 | Part 5 |
|---|---|---|---|---|
01 Overview | 09 Equipment selection & capacity | 14 FAT/SAT strategy | 19 Ramp-up KPI System & Weekly Speed Management | |
02 Cell blueprint & capacity | 10 URS & technical agreement | 15 Equipment manufacturing supervision | 20 Ramp-up Diagnostics: OEE & Root Cause Analysis | |
03 Process route & yield target | 11 Long-lead equipment & logistics | 16 FAT test matrix & acceptance | ||
04 Slurry rheology & electrode | 12 IT/OT, MES backbone arch. | 17 SAT & Line Handover | ||
05 Material & energy balance | 13 BIM & digital twin | 18 Trial Materials, spare parts & change management | ||
06 Layout & compliance | / | / | ||
07 Micro-environment | / | / | ||
08 Safety control loop | / | / |
In the 19 episode, we addressed the ramp-up KPI system and weekly speed management—how to measure whether the factory is accelerating or decelerating. This edition (episode 20) moves from measurement to diagnosis: why OEE alone cannot explain the deep phase of ramp-up, and how a structured root cause framework converts a yield number into an actionable intervention.
Table of Contents
1. The Deep Phase Is a Diagnostic Problem, Not a Performance Problem
Every gigafactory in ramp-up knows its yield is below target. The question is not whether the gap exists. It is whether the organisation can explain the gap with enough precision to close it.
Industry benchmarks show that new production lines during ramp-up may initially operate at OEE levels of 40% or less, while world-class systems achieve more than 85%. This deep phase is not a single failure mode. It is a cluster of interacting losses—availability, performance, and quality—that compound across process steps.
The factories that emerge from the deep phase fastest are not the ones with the best equipment. They are the ones with the most disciplined diagnostic process.
We discussed OEE in episode 19, that OEE alone is not enough, let’s combine it to below.
2. Why OEE Alone Hides the Real Losses
OEE is the standard efficiency metric: Availability × Performance × Quality. It is useful as a headline number, but insufficient as a diagnostic tool.
The reason is structural. OEE aggregates three fundamentally different failure categories into a single percentage, which we discussed before. A low OEE caused by unplanned downtime requires a maintenance response. A low OEE caused by speed loss requires a process engineering response. A low OEE caused by quality loss requires a defect diagnosis response. A factory that tracks only OEE knows it has a problem but not what kind of problem it has.
Traditional OEE also hides losses that are specific to battery manufacturing. Environmental downtime reduces availability but may be classified as planned maintenance, masking its true impact. Yield drift reduces quality but may not trigger a discrete “defect” event. The diagnostic implication is that OEE must be decomposed into its components, and each component must be analysed at the process-step level, not just at the line level.
Then how shall we plan the diagnosis?
3. The Diagnostic Sequence: From Symptom to Mechanism
A structured deep phase diagnostic follows a defined sequence. It does not begin with a meeting. It begins with the data.
Step 1: Decompose the loss. Break OEE into availability, performance, and quality. Identify which component is the dominant contributor to the current gap. If availability is the constraint, the diagnostic focuses on downtime causes. If performance is the constraint, it focuses on speed losses and micro-stops. If quality is the constraint, it focuses on defect mechanisms.
Step 2: Pinpoint the loss. Decompose each component by process step. A line-level availability loss may be concentrated in one or two steps—coating, calendering, winding—while the rest of the line is stable. Pinpoint prevents the common error of applying a factory-wide intervention to a step-specific problem.
Step 3: Stratify the loss. Decompose the pinpointed loss by time, shift, material batch, and equipment state. Is the quality loss concentrated in one shift? One material supplier? One coating machine head? Stratification converts a statistical correlation into a physical hypothesis.
Step 4: Diagnose the mechanism. Apply structured root cause tools—5 Whys, Fishbone (Ishikawa), Pareto analysis—to move from the statistical pattern to the physical mechanism. It is to identify the dominant mechanism that, if corrected, would remove the largest portion of the loss.
Step 5: Verify the intervention. Implement the corrective action, then measure its effect on the specific loss component that was targeted. A corrective action that does not produce a measurable reduction in the targeted loss is not a corrective action.
Below we listed out five major stoppage losses for reference.